A common task for spoken dialogue systems (SDS) is to help users select a suitable option (e.g., flight, hotel, restaurant) from the set of options available. When the number of options is small, they can simply be presented sequentially. However, as the number of options increases, the system must have strategies for helping users browse the space of available options. In this thesis, I compare two approaches to information presentation in SDS: (1) the summarize and refine (SR) approach (Polifroni et al., 2003; Polifroni, 2008) in which the summaries are generated by clustering the options based on attributes that lead to the smallest number of clusters, and (2) the user-model based summarize and refine (UMSR) approach (Demberg, 2005; Demberg and Moore, 2006) which employs a user model to cluster options based on attributes that are relevant to the user and uses coherence markers (e.g., connectives, discourse cues, adverbials) to highlight the trade-offs among the presented items. Prior work has shown that users prefer approaches to information presentation that take the user’s preferences into account (e.g., Komatani et al., 2003;Walker et al., 2004; Demberg and Moore, 2006). However, due to the complexity of building a working end-to-end SDS, these studies employed an ”overhearer” evaluation methodology, in which participants read or listened to pre-prepared dialogues, thus limiting evaluation criteria to users’ perceptions (e.g., informativeness, overview of options, and so on). In order to examine whether users prefer presentations based on UMSR when they were actively interacting with a dialogue system, and to measure the effectiveness and efficiency of the two approaches, I compared them in a Wizard-of-Oz experiment. I found that in terms of both task success and dialogue efficiency the UMSR approach was superior to the SR approach. In addition, I found that users also preferred presentations based on UMSR in the interactive mode. SDS are typically developed for situations in which the user’s hands and eyes are busy. I hypothesized that the benefits of pointing out relationships among options (i.e., trade-offs) in information presentation messages outweighs the costs of processing more complex sentences. To test this hypothesis, I performed two dual task experiments comparing the two approaches to information presentation in terms of their effect on cognitive load. Again, participants performed better with presentations based on the UMSR algorithm in terms of both dialogue efficiency and task success, and I found no detrimental effect on performance of the primary task. Finally, I hypothesized that one of the main reasons why UMSR is more efficient is because it uses coherence markers to highlight relations (e.g., trade-offs) between options and attributes. To test this hypothesis, I performed an eye-tracking experiment in which participants read presentations with and without these linguistic devices, and answered evaluation and comparison questions to measure differences in item recall. In addition, I used reading times to examine comprehension differences between the two information presentation strategies. I found that the linguistic devices used in UMSR indeed facilitated item recall, with no penalty in terms of comprehension cost. Thus, in this thesis I showed that an approach to information presentation that employs a user model and uses linguistic devices such as coherence markers to highlight trade-offs among the presented items improves information browsing. User studies demonstrated that this finding also applies to situations where users are performing another demanding task simultaneously.

The conditional independence assumption imposed by the hidden Markov models (HMMs) makes it difficult to model temporal correlation patterns in human speech. Traditionally, this limitation is circumvented by appending the first and second-order regression coefficients to the observation feature vectors. Although this leads to improved performance in recognition tasks, we argue that a straightforward use of dynamic features in HMMs will result in an inferior model, due to the incorrect handling of dynamic constraints. In this thesis I will show that an HMM can be transformed into a Trajectory-HMM capable of generating smoothed output mean trajectories, by performing a per-utterance normalisation. The resulting model can be trained by either maximisingmodel log-likelihood or minimisingmean generation errors on the training data. To combat the exponential growth of paths in searching, the idea of delayed path merging is proposed and a new time-synchronous decoding algorithm built on the concept of token-passing is designed for use in the recognition task. The Trajectory-HMM brings a new way of sharing knowledge between speech recognition and synthesis components, by tackling both problems in a coherent statistical framework. I evaluated the Trajectory-HMM on two different speech tasks using the speaker-dependent MOCHA-TIMIT database. First as a generative model to recover articulatory features from speech signal, where the Trajectory-HMM was used in a complementary way to the conventional HMM modelling techniques, within a joint Acoustic-Articulatory framework. Experiments indicate that the jointly trained acoustic-articulatory models are more accurate (having a lower Root Mean Square error) than the separately trained ones, and that Trajectory-HMM training results in greater accuracy compared with conventional Baum-Welch parameter updating. In addition, the Root Mean Square (RMS) training objective proves to be consistently better than the Maximum Likelihood objective. However, experiment of the phone recognition task shows that the MLE trained Trajectory-HMM, while retaining attractive properties of being a proper generative model, tends to favour over-smoothed trajectories among competing hypothesises, and does not perform better than a conventional HMM. We use this to build an argument that models giving a better fit on training data may suffer a reduction of discrimination by being too faithful to the training data. Finally, experiments on using triphone models show that increasing modelling detail is an effective way to leverage modelling performance with little added complexity in training.

It has been suggested that theoretical computer science (TCS) suffers more than average from a lack of intrinsic motivation. The reasons provided in the literature include the difficulty of the subject, lack of relevance to the informatics degree or future careers, and lack of enjoyment of the learning experience. This thesis presents evidence of these claims derived from the results of an expert survey. Increasing the students’ perceived control of the learning experience has been shown to increase student motivation in numerous different settings. A few of those also showed increased student performance. This thesis proposes that game-building can be such a setting for the area of TCS. Within the area of TCS, the focus will be on the modelling skills (finite state automata, push-down automata, Turing machines, CCS, etc.) since they form the majority of the curriculum at undergraduate level. It will be demonstrated how arbitrary TCS modelling skills can be mapped onto a game-building framework and allow the students to learn about the former by using the latter. It is hypothesized that the success of the approach depends on the amount of control given to the student. To test this claim, two experimental conditions were used in a repeated-measures design: (1) own-game and (2) pre-defined game. In the former, students are asked to write a game of their own, whereas in the latter, they are asked to copy a pre-defined game. A large demand for the own-game context was observed and results of its effect on performance and enjoyment are presented. Although no main effect of the owngame condition versus the pre-defined game condition was found in terms of either enjoyment or performance, some interesting interaction effects between condition and motivational type were unveiled.

In today’s volatile business environments, collaboration between information systems, both within and across company borders, has become essential to success. An efficient supply chain, for example, requires the collaboration of distributed and heterogeneous systems of multiple companies. Developing such collaborative applications and building the supporting information systems poses several engineering challenges. A key challenge is to manage the ever growing design complexity. In this article, we argue that software architecture should play a more prominent role in the development of collaborative applications. This can help to better manage design complexity by modularizing collaborations and separating concerns. State of the art solutions, however, often lack proper abstractions for modeling collaborations at architectural level or do not reify these abstractions at detailed design and implementation level. Developers, on the other hand, rely on middleware, business process management, and Web services, techniques that mainly focus on low-level infrastructure. To address the problem of managing the design complexity of collaborative applications, we present Macodo. Macodo consists of three complementary parts: (1) a set of abstractions for modeling adaptive collaborations, (2) a set of architectural views, the main contribution of this article, that reify these abstractions at architectural level, and (3) a proof of concept middleware infrastructure that supports the architectural abstractions at design and implementation level. We evaluate the architectural views in a controlled experiment. Results show that the use of Macodo can reduce fault density and design complexity, and improve reuse and productivity. The main contributions of this article are illustrated in a supply chain management case.

No description available.

A recuperação de informação é ainda um assunto essencial a melhorar nos diferentes tipos de sistemas web. Um tipo de sistema web que é muito utilizado na atualidade, é o sistema colaborativo. Estes sistemas permitem que os usuários estejam mais envolvidos, seja contribuindo com a inserção de textos, imagens ou dados, assim como utilizando etiquetas (tags) para identificar aos elementos existentes no sistema e que serão compartilhados com outros usuários. Nesta dissertação utilizamos um sistema colaborativo de compartilhamento de imagens arquitetônicas, onde os usuários podem inserir títulos e tags livremente para descrever uma imagem. Contudo as tags podem ter um significado ambíguo, resultando em imagens recuperadas que não são relevantes, quando são utilizadas técnicas tradicionais, como por exemplo busca booleana ou por palavra-chave. Além disso, os usuários podem utilizar consultas mais complexas utilizando uma linguagem livre, e utilizando as técnicas mencionadas podem recuperar informação não relevante. Assim, esta pesquisa aborda, a construção de uma ontologia no domínio arquitetônico denominada OntoArq, baseada no vocabulário controlado da USP e no tesauro experimental de arquitetura brasileira, a qual possibilitou fortalecer a relação entre as tags e os conceitos estruturados da ontologia, por meio de uso de hierarquias de classes e relações semânticas existentes entre as classes. A ontologia também ajudou a melhorar a recuperação de documentos para consultas complexas que utilizam uma linguagem livre, por meio da adição de termos arquitetônicos relacionados à consulta original dada pelo usuário. E quando a consulta expandida é utilizada em conjunto com o modelo de espaço vetorial existente no sistema de recuperação, auxilia na recuperação de imagens mais relevantes. A avaliação de nossa abordagem foi realizada através de experimentos que utilizaram os dados do sistema Arquigrafia, dois conjuntos de consultas e medidas de avaliação como precisão, cobertura e medida-F. Os conjuntos eram compostos por 11 consultas dada por especialistas da área de arquitetura e 9 consultas aleatórias extraídas do log de busca do Google Analytics do sistema Arquigrafia, tendo um total de 20 consultas. Para nossos experimentos utilizamos as 20 consultas que pertenciam aos dois conjuntos de consultas mencionados, dentre os quais obtivemos resultados positivos para 16 consultas, considerando um valor de precisão, cobertura e medida-F maior do que 50%, com nossa abordagem. Em comparação a outra abordagem, que usa a técnica de busca boolena, obteve-se 1 consulta com resultado positivo, também considerando precisão, cobertura e medida-F maior do que 50%. Assim, podemos concluir que nossa abordagem obteve melhores resultados. Além disso, pelos resultados obtidos, consideramos que nossa abordagem, ao utilizar uma ontologia, pode ser um inicio de como empregar as ontologias como ferramenta de apoio para dar um maior significado semântico às tags que existem num sistema colaborativo e como as ontologias permitem a adição de termos na consulta, sendo estes termos relacionados a uma área do conhecimento, que para nosso caso, a área da arquitetura. Desta maneira podemos recuperar os documentos associados às imagens, os quais serão mais relevantes para consulta feita pelo usuário. / Information retrieval is an essential issue in different web system that needs to be improved. In recent years, collaborative systems have gained popularity. In this thesis, we use a collaborative system of architectural image sharing, that users can add titles and tags freely to describe an image. However, the tags may have an ambiguous meaning, resulting in recovering images that are not relevant when traditional technics are used, such as Boolean and keyword search. In addition, users usually use natural language in their queries and with the mentioned techniques it is less posible to retrieve relevant information. Our research approach is to construct an ontology for architectural domain, based on the controlled vocabulary of USP and the experimental thesaurus architecture information. The ontology allowed to enhance the relationship between tags and structured concepts of ontology, improving the information retrieval through the expanded terms. And when the expanded query is used in conjunction with the vector space model in the recovery system, it helps retrieve more relevant images. The evaluation of our approach was carried out through experiments that used data of Arquigrafia system, two sets of queries were created and metrics like precision, coverage and measure-F were used. The sets consisted of 11 queries given by architecture experts and 9 random queries extracted from Google Analytics logs of the Arquigrafia system, taking a total of 20 queries. In our experiments we used the 20 queries from the two sets, among which we obtained positive results for 16 queries, considering precision, coverage and F-measure values greater than 50%, applying our approach. In comparison to another approach, which uses the Boolean search technique, we obtained 1 query with positive result, also considering the same metrics with values greater than 50%. Thus, we can conclude that our approach has achieved better results. In addition, from the results obtained, we consider that our approach using an ontology can be a beginning of how to use ontologies as a support tool to give a greater semantic meaning to tags that exist in a collaborative system, and how Ontologies allow the addition of semantic terms in queries, like terms related to a knowledge area, in our case the area of architecture. In this way we can retrieve documents associated with images, which will be more relevant to the user\'s query.

Na Web 2.0 são encontrados sistemas com alto volume de interação social. Alguns desses sistemas oferecem cálculo de reputação ou alguma forma de classificação de usuários ou do conteúdo compartilhado. Contudo, em muitos casos, esse valor de reputação resultante é obtido somente a partir de dados quantitativos ou qualitativos. O objetivo deste trabalho é elaborar um modelo para o cálculo de reputação em comunidades on-line, baseando-se em dados qualitativos e quantitativos provenientes da interação dos próprios participantes da rede, a fim de potencializar a colaboração entre os membros e fornecer um meio de cálculo resistente a algumas das vulnerabilidades comuns em sistemas de reputação, como tolerância a ruídos e ataques Sybil. Para atingir esse objetivo é realizada uma adaptação do algoritmo PageRank, definida como CR (Collaborative Reputation) para obter uma ordenação dos usuários a partir de suas interações. Para avaliação, adotamos um conjunto de dados do sítio Epinions.com, com o qual foi realizada uma análise comparativa dos resultados obtidos a partir do modelo proposto com outros três algoritmos correlatos ao trabalho apresentado. Dentre as técnicas usadas na análise estão: diversidade de valores, comparação da ordenação, estudo comparativo de cenários, tolerância a ruídos e robustez contra ataques tipo Sybil. Os algoritmos usados na avaliação são: o PageRank original e o algoritmo ReCop, usados para a identificação de usuários relevantes, e o algoritmo LeaderRank usado para a identificação dos usuários com maior prestígio na rede. Os resultados indicam que o modelo proposto é mais sensível às interações dos usuários em comparação aos outros modelos usados na avaliação, mas é mais eficiente a ataques Sybil. / In the Web 2.0, there are systems with high volume of social interaction. Some of these systems offer reputation calculation or some form of classification of users or shared content,. However, in many cases, this reputation value is obtained solely from quantitative or qualitative data. The objective of this work is to develop a model for the reputation calculation in online communities, based on qualitative and quantitative data from the interaction of the participants of a social network, in order to potentiate the collaboration between users, and to provide a resistant environment for some of the vulnerabilities present in reputation systems. To achieve this goal we defined an adaptation of the PageRank algorithm, defined as CR (Collaborative Reputation), to obtain a rank of users based on their interactions in the network. For evaluation, we used a dataset from the site Epinions.com. With that database, we executed a comparative analysis of the results of the proposed algorithm and of three other algorithms related to the presented work. The procedures used in the analysis were: diversity of values, comparison of ordination, comparative study of scenarios, noise tolerance and robustness against Sybil attacks. The algorithms used in the comparison were: the original PageRank algorithm and ReCop, used to identify relevant users, and the algorithm LeaderRank, which is used for identification of the most prestigious users in the network. The results showed that the proposed model is more sensitive to the interactions of users, but its performance on Sybil attacks is better than the others.

With today’s smartphones and smart glasses, and the progression of augmented reality, the possibilities to interact over distance has made it feasible to guide one another in an intuitive and effective way. This combination of technology and software principles allows a local operator to record a scenario from his or her point of view and show this to a remotely located expert. An expert can, in turn, help the operator by interacting through that video feed by highlighting aspects and overlaying information for the operator to see. The aim of this study was to investigate how such a system should be configured if it were to be implemented in the context of final assembly of military aircraft. An understanding of the context and situations where external help might be needed was established through an ethnographical study. User tests were conducted with an existing system in comparable cases, inspired by results from the first study. This in order to evaluate the configuration of the hardware and interactivity. Results indicate that it is useful to implement a remote guidance system which allows augmented overlays in the context of final assembly. A greater need for such a system was found in situations in which a subject matter expert must investigate and assess issues and errors that has occurred. These scenarios are often characterized by differentiating environment, from cases where there is a good overview to situations in which the mirrors must be used to see from the right angle beyond one’s field of view. A remote guidance system should be able to support both cases and must, therefore, be modular in a way so that an external camera can be used to reach in while the screen can be seen simultaneously. Regarding the need for interaction between interactors in such situations are limited. The user studies indicate that simple referential gestures on frozen images of a video feed might be enough. / Global Assembly Instruction Strategies (GAIS) 2

Submitted by Johnny Rodrigues (johnnyrodrigues@ufcg.edu.br) on 2018-05-08T16:56:51Z No. of bitstreams: 1 FELIPE VIEIRA FALCÃO - DISSERTAÇÃO PPGCC 2015..pdf: 16387085 bytes, checksum: 904dfed8e6bc60733d727942a3884611 (MD5) / Made available in DSpace on 2018-05-08T16:56:51Z (GMT). No. of bitstreams: 1 FELIPE VIEIRA FALCÃO - DISSERTAÇÃO PPGCC 2015..pdf: 16387085 bytes, checksum: 904dfed8e6bc60733d727942a3884611 (MD5) Previous issue date: 2015-12-01 / Capes / O ato de ouvir música socialmente em conjunto com outros indivíduos é uma experiência frutífera e cada vez mais comum no dia a dia das pessoas. Social jukeboxes são sistemas que permitem que usuários realizem tal atividade por intermédio da escolha colaborativa das faixas a serem executadas. Naturalmente, como os gostos musicais dos diferentes usuários são diversos, o uso de social jukeboxes geralmente está atrelado à presença de conflitos. Por conta disso, todas as social jukeboxes desenvolvidas para ambientes virtuais incorporam ao seu funcionamento mecanismos de gerenciamento de conflitos. Em contraste com suas amplas utilizações, muito pouca atenção tem sido dada no que se refere à avaliação de como os mais diferentes mecanismos de gerenciamento de conflitos atuam em favor da preservação da experiência positiva dos usuários no que se refere à manutenção de altos níveis de satisfação. Esta pesquisa apresenta um experimento composto da observação de três mecanismos de gerenciamento de conflitos aplicados em três grupos distintos de usuários por intermédio de uma social jukebox on-line. Os mecanismos foram escolhidos de maneira a representar aqueles mais frequentemente utilizados em aplicações presentes no mercado e na literatura. Nosso estudo emprega uma abordagem de métodos mistos para analisar quantitativamente a satisfação dos usuários ao passo que também examina suas impressões e visões sobre conflitos, mecanismos de gerenciamento dos mesmos e a atividade de social jukeboxing. Tais abordagens metodológicas convergiram para conclusões relacionadas à efetividade dos três mecanismos escolhidos para estudo além da notável superioridade do mecanismo de upvoting/downvoting em todos os grupos observados. Entrevistas listaram vários exemplos de conflitos atrelados à utilização do tipo de ferramenta proposta em conjunto com sugestões sobre como se gerenciam tais ocorrências. / Social music listening is a prevalent and often fruitful experience. Social jukeboxes are systems that enable social music listening with listeners collaboratively choosing the music to beplayed. Naturally, because music tastes are diverse, using social jukeboxes often involves conflicting interests. Because of that, virtually all social jukeboxes incorporate conflict management mechanisms. In contrast with their widespread use, however, little attention has been given to evaluating how different conflict management mechanisms function to preserve the positive experience of music listeners. This research presents an experiment with three conflict management mechanisms and three groups of listeners. The mechanisms were chosen to represent those most commonly used in the state of the practice. Our study employs a mixed-methods approach to quantitatively analyze listeners’ satisfaction and to examine their impressions and views on conflict, conflict management mechanisms, and social jukeboxing.

Submitted by Geandra Rodrigues (geandrar@gmail.com) on 2018-01-26T10:38:41Z No. of bitstreams: 0 / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2018-01-26T10:56:19Z (GMT) No. of bitstreams: 0 / Made available in DSpace on 2018-01-26T10:56:19Z (GMT). No. of bitstreams: 0 Previous issue date: 2015-09-30 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Fornecer suporte a interoperabilidade apenas considerando a forma e o significado (i.e. sintaxe e semântica) na troca de dados não é suficiente para se atingir uma colaboração efetiva e significativa. Neste sentido, a interoperabilidade pragmática tem se destacado como um requisito fundamental para garantir a colaboração em sistemas distribuídos. Entretanto, preencher este requisito não é uma tarefa trivial. O objetivo deste estudo é propor e avaliar uma solução para apoiar implementação da interoperabilidade pragmática em um sistema colaborativo. Assim, a solução proposta foi implementada e avaliada em um ecossistema de software baseado na web capaz de apoiar o desenvolvimento colaborativo de workflows científicos chamado ECOS Collaborative PL-Science. / Providing interoperability support only considering the format and meaning (i.e. syn-tax and semantic) in data exchange is not enough to achieve effective and meaningful collaboration. Pragmatic interoperability has been identified as a key requirement to fos-ter collaboration in a distributed environment. However, fulfilling this requirement is not a trivial task. The aim of this study is to propose and evaluate a solution to support pragmatic interoperability implementation in a collaborative system. The proposed solu-tion was implemented and evaluated in an open source web-based software ecosystem to support collaborative development of scientific workflows.